A fleet dashboard is worth building when you run three or more skylight cleaning robots. Below that, a spreadsheet is faster. The metrics that matter in 2026 are cleaning metres, error codes, battery health and water consumption per square metre.
Facilities groups with robot fleets usually start with a whiteboard, then a spreadsheet, then a mess. The jump to a dashboard is not about technology, it is about catching a failing machine before it fails on a client site. That is the whole point.
Which metrics actually help?
- Cleaning area per run and per day, in m2.
- Error codes, grouped by type not by timestamp.
- Battery cycles and depth of discharge.
- Water used per m2, which flags a clogged nozzle early.
- Service dates for cups, brushes and seals.
Four of those five you will get from the machine logs. The fifth, service dates, you have to track yourself, and it is the one that saves the most money because it turns maintenance from reactive into scheduled.
What should you ignore?
Live GPS maps of robots moving across a roof are impressive in a demo and useless in practice. Distance travelled is noise. Hours of robot uptime is noise too, unless you bill per hour, which almost nobody does for glass cleaning. Focus on output per unit of input and on errors that repeat.
One genuinely useful proxy: if water per m2 climbs 20% over its rolling average, you have a nozzle or filter issue, and you will find it before a client complains about streaks. That single alert is worth the dashboard effort.
How do you connect the machines?
Not every robot exposes an API, and that is normal. If your machines log to a USB stick or a local app, set a weekly export and a simple script to append it to a database. You do not need live streaming. Weekly is enough to catch trends, and it avoids the trap of a fancy dashboard nobody updates.
If you run mixed fleets, standardise one export format. Teams that deploy Lingkong K3 alongside older units usually build the format around the newer machine and map the older data into it. Set that standard before you buy unit number four, not after. We help teams plan fleet rollout — see the contact page.
Key Takeaways
- A dashboard pays for itself at three or more robots.
- Track m2 cleaned, error codes, battery health, water per m2.
- Skip GPS maps and uptime hours, they are noise.
- A 20% rise in water per m2 signals a nozzle or filter fault.
- Set one export format before buying a fourth robot.